794
J. Phys. Chem. A 2003, 107, 794-803
Electron Autodetachment from Isolated Nickel and Copper Phthalocyanine-Tetrasulfonate Tetraanions: Isomer Specific Rates Katharina Arnold,† Teodor Silviu Balaban,† Martine N. Blom,† Oli T. Ehrler,‡ Stefan Gilb,‡ Oliver Hampe,† Johan E. van Lier,§ J. Mathias Weber,‡ and Manfred M. Kappes*,†,‡ Institut fu¨ r Nanotechnologie, Forschungszentrum Karlsruhe, P.O. Box 3640, D-76021 Karlsruhe, Germany, Institut fu¨ r Physikalische Chemie, UniVersita¨ t Karlsruhe, Kaiserstr. 12, D-76128 Karlsruhe, Germany, and Faculty of Medicine, UniVersity of Sherbrooke, Sherbrooke, Que´ bec, Canada ReceiVed: September 25, 2002; In Final Form: December 2, 2002
Electron autodetachment from isolated metal phthalocyanine-tetrasulfonate tetraanions (MPc(SO3)44-) (M)Ni, Cu) was studied at room temperature in an FT-ICR mass spectrometer under UHV conditions. For M ) Cu, the electron loss rate was measured for the as-prepared isomer mixture. In the case of M ) Ni, two chromatographically purified fractions containing different compositions of constitutional isomers were studied. The observed rate constants for these fractions differed by a factor of 37 ( 5. Photoelectron spectroscopy indicates that both Cu and Ni compounds comprise isomer ensembles which are electronically metastable. Model calculations suggest that the differing electron loss rates observed for the Ni isomer fractions reflect tunneling through repulsive Coulomb barrier surfaces, which are themselves strongly isomer dependent in shape and height.
1. Introduction Small multiply charged molecular anions (MCAs ) multianions) are quite fragile and often metastable in the gas phase. The strong Coulombic repulsion between close-lying negative charges can result in spontaneous fragmentation, electron loss, or both. Nevertheless, as increasingly sophisticated experimental tools have become available, smaller and smaller multiply charged anions such Cn2- (n g 7),1 S2O62-,2 SO4(H2O)n2(n g 3),3 or MX42- (M ) Pt, Pd; X ) Br, Cl)4,5 have become accessible for gas phase studies. This has the added advantage of providing experimental benchmarks for small computationally tractable molecules against which evolving theoretical descriptions of multianion dynamics can be tested. In particular, the recent photoelectron spectroscopic measurements of Wang et al. have allowed fundamental insights into the characteristic electron detachment phenomena associated with multianions.4,6 PtCl42- is an example of a metastable dianion with a negative electron detachment energy (i.e., electronically unstable). In this case, spontaneous electron autodetachment may be observed on a time scale of seconds at room temperature.5 This requires that the outgoing electron surmounts a nonlocal Coulomb barrier potential surface which is common, in a qualitative sense, to the ground and the excited states of all multianions. In principle, there are two mechanistic limits in which to think about this: (i) thermal electron emission (thermoemission) oVer the Coulomb barrier and (ii) tunneling through the Coulomb barrier, for the latter, a negative detachment energy is a necessary criterion. In the case of thermal emission,7 rate constants can be modeled using the unimolecular rate theory developed for * To whom correspondence should
[email protected]. † Forschungszentrum Karlsruhe. ‡ Universita ¨ t Karlsruhe. § University of Sherbrooke.
be
addressed.
E-mail:
thermoionization of neutral molecules (Klots theory) which is essentially dependent on the assumed magnitude of the Coulomb barrier, the size of the molecule, and the level of vibrational excitation. As first suggested by Landman et al.8 for electronically unstable multianions, tunneling through a Coulomb barrier surface can be roughly described by the Wentzel-KramersBrillouin (WKB) approximation using the Coulomb barrier properties and the electron detachment energy. If both mechanisms are energetically possible, it is not clear a priori which one dominates. Consequently, in studying autodetachment, it is of interest to vary the excitation level. In our recent study of electron autodetachment from PtCl42-, we have demonstrated that the level of vibrational excitation at room temperature is insufficient to allow for any thermoemission and that, therefore, the observed electron loss must be entirely attributable to electron tunneling. Spontaneous electron loss from metastable multiply charged ions has also been theoretically investigated.9,10 Both accurate initial and final state energies as well as aqequate descriptions of Coulomb barriers are crucial in order to quantitatively explain the experimentally observed electron autodetachment rates. Among other issues, the question of the localization/delocalization of the excess charges and the corresponding influence on the topology of the Coulomb barrier (e.g., 2D versus 3D) surface arises. Atomic clusters, which can be experimentally probed over a wide range of sizes and bonding types, are quite useful in this context. Examples of cluster MCAs with (partially) delocalized charge distributions, which have recently been studied experimentally, include (i) multiply charged metal cluster anions such as Agn2- (n > 28)11 or Pbn3- (n > 76)12 and (ii) fullerene dianions Cn2- (n ) 60)13,14 and (n g 70).7,15 Most experimentally known molecular MCAs may however be described to first order by localized charges, typically at the periphery of the molecule. Examples include three-dimensional alkali halide cluster multianions16 or (quasi) one-dimensional aliphatic dicarboxylates -OOC-(CH2)n-COO- (n ) 3-10).17
10.1021/jp022141o CCC: $25.00 © 2003 American Chemical Society Published on Web 01/18/2003
Isolated Nickel and Copper MPc(SO3)44Correspondingly, first-order properties, like vertical detachment energies, are in accordance with simple electrostatic calculations employing the appropriate localized charge distributions. Excepting barrier heights, which have been roughly bracketed by comparing photoelectron spectra for different detachment energies, little is known experimentally about the Coulomb barriers themselves. On a theoretical basis, they are expected to be spatially anisotropic, nonlocal, as well as dependent on both the initial and final states. A first attempt toward a more detailed characterization of barrier “surfaces” was described in a recent PES study of the three possible isomers (ortho-, meta-, and para-) of benzene dicarboxylate C6H4(COO-)2, which are electronically stable dianions having the charges localized on the carboxylate group.18 In this case, the measurements revealed a shift both in detachment energies and barrier heights due to the different intramolecular Coulomb repulsions depending on the relative position of the extra charges within the molecule.18 One may thus infer that analogous electronically metastable isomers might have measurably different electron autodetachment rates. In this study, we show, by using a set of electronically metastable tetraanions as model systems that such variations indeed occur and that they can in fact be quite large. Wang et al. have recently reported the highest known negative electron detachment energy of -0.9 eV in measurements of isolated copper phthalocyanine-tetrasulfonate tetraanion.19,20 On the basis of our previous studies, such systems might then be expected to undergo measurable tunneling autodetachment at room temperature. Furthermore, it is known that these and related transition metal phthalocyanine-tetrasulfonates (MPc) are synthesized by a procedure which generates a mixture of several constitutional isomeric forms which differ only in the relative positions of their sulfonic groups. These groups account for the excess charges and are mounted on the molecular perimeters. No previous separations of these constitutional isomers have so far been reported. Consequently, we set out to study the electron autodetachment dynamics of two readily available MPc(SO3)44- systems: copper- and nickel-phthalocyanine-tetrasulfonate tetraanions. Both were obtained as the sodium salts and were studied in gas phase via electrospray volatilization from the appropriate polar solutions. In the case of CuPc(SO3)44-, we probed only the isomer mixture. In the case of NiPc(SO3)44-, which is accessible to proton-NMR characterization, we succeeded in partial isomer separation and were then able to determine autodetachment rates of selected constitutional isomer fractions. Specifically, this contribution comprises (i) preparation and partial separation using high-pressure liquid chromatography (HPLC) of constitutional isomers of nickel and copper phthalocyanine-tetrasulfonate NiPc(SO3)44- and CuPc(SO3)44- and in the case of NiPc their characterization and assignment using NMR spectroscopy in solution; (ii) electrospray ionization/ volatilization of the corresponding tetraanions as well as their high-resolution mass spectroscopic analysis; (iii) determination of photodetachment energies and photoelectron spectra for isomer mixtures of both M ) Ni and Cu; (iv) measurements of the room temperature autodetachment rates for the nickel pthalocyanine tetraanion fractions obtained; and last (v) a semiempirical quantum chemical as well as classical electrostatic modeling study of the electron loss kinetics. 2. Methods 2.1. Sample Preparation and Characterization. Copper and nickel phthalocyanine-tetrasulfonate were prepared via the condensation method adapted from Weber and Bush.21 Briefly this involved the following.
J. Phys. Chem. A, Vol. 107, No. 6, 2003 795 Nickel-Phthalocyanine-Tetrasulfonate Sodium Salt (MW 978). A mixture of nickelous chloride hexahydrate (MW 237.7, 4.4 g, 18.4 mmol), triammonium sulfophthalic acid (MW 297, 10.87 g, 36.6 mmol), ammonium chloride (MW 53, 0.35 g, 6.5 mmol), ammonium molybdate (0.1 g), and urea (MW 60, 10 g, 0.16 mmol) was heated 3 h at 200 °C, triturated with 1N HCl in brine, filtered and washed with water, dialyzed, evaporated to dryness, and oven-dried to give 7.16 g (80% yield) of the nickel phthalocyanine tetrasulfonate. λmax in DMF: 669 nm (max ) 155 000 L mol-1 cm-1). Copper-Phthalocyanine-Tetrasulfonate Sodium Salt (MW 983.5). A mixture of copper diacetate monohydrate (MW 200, 8 g, 40 mmol), triammonium sulfophthalic acid (MW 297, 39.92 g, 134 mmol), ammonium chloride (MW 53, 8 g), ammonium molybdate (40 mg), urea (MW 60, 40 g, 0.64 mmol), and water (10 mL) was heated 4 h at 190-200 °C, triturated with 1N HCl in brine, filtered and washed with water, dialyzed, evaporated to dryness, and oven-dried to give 18.32 g (55% yield) of the copper phthalocyanine tetrasulfonate. λmax in MeOH: 666 nm. Preparative HPLC separations were performed by means of a Varian ProStar setup equipped with a UV-vis detector and two Rainin pumps on a reverse phase Dynamax Microsorb C18 column of 250 mm length and 4.6 mm internal diameter which was at room temperature. Solvents (water and methanol) were HPLC grade, and the buffer was freshly prepared and the pH adjusted with sodium dihydrogen phosphate (Merck) to be between 5.1 and 5.30 as measured by a Mettler-Toledo MP 225 pH meter. Mixing and degassing of the eluent (25% methanol in aqueous buffer or in pure water) was effected before it was passed through the column which was protected by a Microsorb C18 guard module. The flow rate was usually 20 mL/min. NMR spectra were recorded overnight on an Bruker Avance 300 spectrometer (300 MHz for protons) in DMSO-D2O solutions (2:1 to 3:1 vol/vol) where the aggregation was inhibited as observed from absorption spectra measured on Varian Cary 500 instrument. The temperature was set at 75 °C and controlled by the variable temperature unit. The HOD peak was irradiated (PRESAT pulse sequence) in the undecoupled spectra. 2.2. Mass Spectrometry and Ion Trapping. All measurements of autodetachment kinetics were performed with a 7 T Fourier transform ion cyclotron resonance mass spectrometer FT-ICR-MS (APEX-II, Bruker Daltonics) equipped with an electrospray ion source (Analytica of Branford). The ions were obtained upon spraying a 0.1 mmol aqueous solution of the tetrasodium salt of NiPc(SO3)44- and CuPc(SO3)44-, respectively, at a typical spraying rate of 100 µL/h. The source contains a hexapole (pre)trap storing the ions for 1 s at 10-4 mbar. For that reason, the ions are estimated to be thermalized to room temperature. Experimental details have been described recentl.5 The Penning trap was held at room temperature and at a pressure of 2 × 10-10 mbar. To acquire unimolecular rates, we chose the following experimental sequence: 1. Trapping: ions were trapped “on the fly” and stored in a electrostatic potential well of typically 1.0-1.5 V depth. 2. Isolation: besides the 4-fold charged species of interest, all other ions were removed from the trap by means of resonant rf excitation. 3. Reaction delay: during a period of up to 1000 s, the ions were allowed to undergo unimolecular decay reactions, in this case electron loss. By varying the length of this delay the kinetics of the reaction can be studied. 4. Detection and quench: all ions, precursor and products, were detected after broadband rf excitation and then removed
796 J. Phys. Chem. A, Vol. 107, No. 6, 2003
Arnold et al.
SCHEME 1
from the trap by pulsing the potential applied to the trapping electrodes of the ICR cell. This sequence was typically repeated 32 or 64 times, and the data were accumulated in order to achieve a good signalto-noise ratio. Kinetic data were taken and processed as recently described.22 Briefly, detected ICR signals from product and educt ions were corrected for their charge state to give relative ion intensities.23 For single exponential behavior, a unimolecular rate constant k can be determined by fitting a straight line to the logarithmically plotted data, corresponding to the equation
ln(It/It)0) ) -kt
(1)
where It and It)0 denote the intensities of the tetraanion at time t and time 0, respectively. 2.3. Photoelectron Spectroscopy. As-prepared isomer mixtures of CuPc(SO3)44- and NiPc(SO3)44- were probed using a newly constructed magnetic bottle photoelectron spectrometer after Kruit and Read24 coupled to a reflectron time-of-flight mass spectrometer (RETOF-MS) equipped with an electrospray source and a hexapole trap (Analytica of Branford) as described above. Details of the machine, which is based on a design by Cheshnovsky et al.25 will be reported elsewhere. Briefly, 0.1 mmol solutions of the respective tetrasodium salts in a mixture of methanol and water (98:2 vol %) were electrosprayed from an off-axis microspray needle using a solution flow rate of 10 µL/min and a drying gas (N2) temperature of 200 °C. The resulting ions were accumulated for 1/30 s in a hexapole trap at 10-4 mbar. As the main gas load in this region stems from the drying gas, we expect the ions to be above room temperature. At the end of each trap period, the ion bunch was extracted and focused into the injection region of the RETOF-MS by an electrostatic lens system. At an appropriate time delay after the trap extraction pulse, ions were accelerated at a starting potential of 485 V in a RETOF-MS perpendicular to the primary ion beam, decelerated to ∼30 eV, and irradiated using the third and fourth harmonic of a Nd:YAG laser (Spectra Physics, Quanta RAY series, LAB 150-30). Photoelectrons were collected at nearly 100% efficiency using the magnetic mirror effect in an inhomogeneous magnetic field and allowed to disperse according to their kinetic energy in a 1.68 m drift tube with a weak guiding magnetic field. Electron time-of-flight spectra were acquired by amplifying the electron signal from a double micro channel plate detector in a fast preamplifier and sampling events with a multichannel scaler. The energy axis was calibrated against the well-known photoelectron spectra of I- at 355 and 266 nm (corresponding to photon energies of 3.49 and 4.66 eV, respectively). The spectra were accumulated over 3 × 105 pulses and corrected for background events due to multiphoton ionization of background gas, electronic noise, etc. 2.4. Computation. Quantum mechanical computations were performed using HyperChem (Release 6.0).26 First, we optimized the ground-state geometric structure of the NiPc(SO3)44-
tetraanions for all four isomers at a semiempirical level (PM3). PM3-Hamiltonians have been shown to give good descriptions of structural properties of analogous organic as well as organometallic systems.27 To obtain a better description of the electronic structure, these PM3 geometry optimizations were followed by a single-point calculation at the ZINDO/S, restricted Hartree-Fock level. This INDO (intermediate neglect of differential overlap) method has been parametrized28 to reproduce electronic properties (like orbital energies and UV-vis spectra) in comparable systems. The tetraanions were calculated as closed shell spin singlets, whereas trianions are expected to have one unpaired electron (spin doublet). The vertical electron detachment energies were taken as the differences between the restricted HF total energy of the quadruply and triply charged ion (at ZINDO/S level), both taken at the optimized geometric structure of the 4-fold charged species. Further classical electrostatic modeling was performed using the software package Mathematica (version 4.0). 3. Results and Discussion 3.1. Sample Preparation and Isomer Separation. Two methods exist for preparing polysulfonated phthalocyanines (MPc(SO3)nn-): (i) direct sulfonation of a Pc metal complex and (ii) condensation of adequately functionalized derivative of phthalic acid, such as the dinitrile, diamide, anhydride, or diiminoisoindoline also bearing a sulfonated group (usually as the sodium salt) in the presence of a metal salt and base. The latter reaction is depicted in Scheme 1 for the case of the dinitrile bearing a 4-substituent (R ) SO3-Na+). The percentages listed beside the symmetry point group, to which each isomer belongs, indicate the expectation for the statistical formation of the four possible isomers. It is generally thought that the templating metal ion has a strong influence upon the isomeric distribution which may deviate significantly from the statistical outcome. Also noteworthy is that the condensation method using only one dinitrile precursor yields fewer isomers than are obtained by the direct sulfonation of MPc in which case both the 3- and 4-positions of each c-benzo-annulated pyrrole ring become monosubstituted. This was proven by a careful inspection of the rather complex HPLC traces of reaction mixtures obtained by both methods.29,30 In all previous HPLC studies known to us, the MPc(SO3)44- isomers were always eluted as essentially one peak with a very low retention time, before the corresponding MPc(SO3)33- or MPc(SO3)22- for which an adequate separation of isomers could be achieved.29 In this study, we were able to devise a protocol to improve the chromatographic separation of both CuPc(SO3)44- and NiPc(SO3)44- isomeric mixtures obtained by the condensation method. By carefully adjusting the pH (between 5.1 and 5.3) of the phosphate buffer (Na2HPO4/NaH2PO4) used in conjunction with methanol to elute the isomeric MPc(SO3)44- mixture
Isolated Nickel and Copper MPc(SO3)44-
J. Phys. Chem. A, Vol. 107, No. 6, 2003 797
Figure 1. HPLC traces obtained upon applying the protocol (as described in the text) to an isomeric mixture of NiPc(SO3)4Na4.
from a reverse phase C18 preparative column, partial separation could be achieved into a rapidly (i.e., with a retention time under seven minutes) eluting cluster of four peaks. In addition to this initial elution, we observed that more MPc(SO3)44- material may be eluted isomer specifically by passing buffer through the column for an extensive time period (up to 120 min) and then rapidly switching the eluent from buffer/methanol to water/ methanol (i.e. from pH 5.5 to pH 7.0). Subsequently, a second cluster of peaks (labeled A′, B′, C′, and D′ in Figure 1) is detected and various phthalocyaninic fractions can be collected. The separation of these later eluted peaks is dependent on the pH of the buffer, on the buffer washing time, and on the methanol:buffer, or methanol:water ratio of the eluent. This makes an optimization of such a multiparametric separation tedious. We could not achieve baseline separation for all four isomers in the final cluster of four peaks. However, one component (fraction A′), which according to the 1H NMR for the NiPc(SO3)44- isomeric mixture is the main component (about 50%), could be obtained in a purity over 95%. Similarly, from the cluster of early buffer-eluted peaks, one component was obtained in almost pure form again according to 1H NMR. We cannot exclude that, beside the usual adsorption/desorption equilibria on the nonpolar stationary phase, two other equilibria, namely, aggregation/deaggregation and counterion pairing or protonation (i.e., SO3-Na+/SO3H), occur concomitantly, during the pH ramping phase from the buffer to water eluents. Evidence for strong aggregation in such solutions comes both from UV-vis31 and 1H NMR spectroscopic studies. Conditions where the formation of dimers and higher aggregates is largely inhibited, thus allowing finely resolved 1H NMR spectra to be recorded, are the following: dilute (i.e., 10-100 µmolar concentrations) in a D2O/DMSO-d6 solution (1:2 to 1:3, vol: vol) and at 75 °C. Because of the low concentrations, usually overnight accumulation of the proton spectra, using a solvent suppression pulse sequence (PRESAT, with irradiation of the residual HOD signal) had to be employed, to achieve a satisfactory signal-to-noise ratio which allowed the assignment of signals, as indicated in Figure 2a. Although in water the H aggregate is predominant, addition of DMSO and increased temperature clearly lead to monomerization. A detailed 1H NMR study including 2D-COSY and homodecoupling experiments was undertaken with the first eluted fraction after changing the eluent (A′ in Figure 1). The latter spectra are shown in Figure 2b and allow an unambiguous assignment of this component, isolated in about 95% purity, to the C2V isomer. Thus, the high field doublet at 8.55 ppm, with a 6.6 Hz splitting, is assigned to protons situated in β position to the sulfonic group, whereas the two low field signals must
Figure 2. 1H NMR spectra of NiPc(SO3)4Na4 solutions: (a) Isomer mixture (upper trace), HPLC-separated fractions B + C (middle trace) and A′ + B′ (lower trace). Note that fine splittings are not observed due to Presat pulse sequence; (b) Decoupling 1H NMR spectra of fraction A′ + B′ without irradiation (lower trace) and irradiating proton HA (middle trace) and HB (upper trace).
be in adjacent R positions. As the 9.40 ppm signal is much broader, here the 6.6 Hz coupling must be operative so this signal can be assigned to the proton adjacent both to the sulfonic group and the other 8.55 ppm resonating proton. The sharper, lowest field proton (at 9.64 ppm) is thus the proton between the sulfonic and quaternary phthalocyaninic carbon, in accord with simple estimations of chemical shifts of a trisubstituted benzo-annulated ring-current bearing aromatic moiety. The large 6.6 Hz vicinal coupling is clearly seen in the H-H COSY spectrum from the cross-peak which correlates the 8.55 ppm signal to the 9.40 ppm signal. Revealingly, upon irradiation of the high field doublet, while the 9.4 signal sharpens, two distinct singlets, of equal intensity, are obtained for the low field signal. On the other hand, upon irradiation of the latter signal which leads to cancellation of the small J4 and J5 couplings, the broad 9.4 Hz signal becomes a triplet (i.e., two doublets with 7.8 and 8.1 Hz splitting having the middle line superimposing), whereas the high field doublet becomes sharp, showing a 7.8 Hz splitting. This proves that, although the high field protons are in the (almost) same magnetic environment at 300 MHz, two equal populations exist for the other two protons which are adjacent to the sulfonic group. Only the C2V isomer fulfills this requirement. An alternative interpretation could be that this fraction consists of a 1:1 mixture of the D2h and C4h isomers. However, this is a much less probable case.
798 J. Phys. Chem. A, Vol. 107, No. 6, 2003
Arnold et al.
Figure 4. Comparison of experimentally observed electron autodetachment rates for copper and nickel phthalocyanine-tetrasulfonatetetraanions (isomer ensembles). Rate constants and half-lives are from a linear fit (assuming a unimolecular kinetics) to the data points on the semilogarithmic plot.
Figure 3. Experimental sequence used to take kinetic data for electron auto-detachment from isolated NiPc(SO3)44-: (a) Negative ion ESIFT-ICR mass spectrum of Ni-phthalocyanine-tetrasulfonate solution. (b) Same as in part a after ejection of all unwanted ions from the trap by means of resonant RF excitation. (c) Mass spectrum taken after a set delay time under UHV.
3.2. Gas-Phase Studies: Mass Spectrometric Analysis and Metastable Decay of MPc(SO3)44-. Figure 3a displays a typical negative ion FT-ICR mass spectrum upon spraying an aqueous solution of the isomer mixture of Na4[NiPc(SO3)4]. The strongest ICR signal stems from the 4-fold negatively charged NiPc(SO3)44around m/z 223. Isotopically resolved mass peaks (not shown here) prove the charge state of all ions described. We also detect NiPc(SO3)4H3- and NiPc(SO3)4Na3-, as well as small amounts of fragments, e.g., NiPc(SO3)33-. Solutions of various isomer fractions (separated by HPLC, see above) were sprayed as obtained from the column and show virtually the same negative ion distributions over the pH range (5.5-7) covered in this study. For fractions C′ and D′, we observe a slightly increased relative intensity of the NiPc(SO3)33-peak. Mass analysis of the CuPc-tetrasulfonate solutions revealed the same relative ion intensities, the CuPc(SO3)44- (m/z ) 224) signal being the most abundant. To study the stability of the observed multiply charged anions, the ion of interest is unambiguously isolated in the ICR cell as described in section 2.2. Although the triply charged species are stable under UHV conditions (10-10 mbar), the 4-fold charged anion undergoes metastable decay. For NiPc(SO3)44-, the only decay channel observed at room temperature is electron autodetachment
NiPc(SO3)44- f NiPc(SO3)43- + e-
(2)
This is evidenced in Figure 3c showing the mass spectrum after isolating the tetraanion in the trap and a reaction delay of 150 s. Note that we can rule out a significant contribution of collision-induced electron loss for two reasons: (i) autodetachment studies were always performed at nominal pressures of
less than or equal to 10-10 mbar and (ii) intentional addition of argon collision gas to the trap at pressures of ca. 10-9 mbar gives rise to primarily fragment ion formation under otherwise identical conditions. No such fragmentation was observed under UHV. Interestingly, electron autodetachment is also observed for CuPc(SO3)44-. Figure 4 contrasts the electron autodetachment kinetics of both metallophthalo-cyanine tetraanions in a semilogarithmic plot. Although the data do not exhibit a single exponential behavior, a fact which we interpret is due to the mixture of isomers which are present in the samples in both cases, we formally assign an average half-life to both data sets by a linear fitting procedure. Interestingly, the half-lives differ only slightly as a function of the central metal atom (240s vs 275s for Ni and Cu, respectively). This is consistent with our computational finding (see also next section) that the highest occupied molecular orbital (HOMO) of the NiPc tetraanion essentially resides on the organic moiety with very little contribution from the chelated metal atom. In this context, it is interesting to invoke photoelectron spectra of neutral NiPc and CuPc,32 where it was shown that lowest energy ionization occurs from the a1u orbital of the Pc ring again with only small variation (
